|
所在平台: Udemy |
课程主页: https://www.udemy.com/course/design-of-experiments-doe-f/
课程评论：没有评论
Coursera 上的《实验设计（DOE）》课程，旨在教授学员如何通过系统性的实验方法来优化产品或流程。 课程以一个披萨店的例子生动地引入了实验设计（DOE）的概念和优势。假设你是一名披萨店老板，希望做出您所在城镇中最美味的披萨。披萨的美味程度受到多种因素影响，包括面团（面粉、酵母、盐）、馅料（酱料、奶酪）以及烤箱温度、烘烤时间等。 课程将重点关注披萨面团的配方，并将面团的三个关键因素设为实验变量： * **x1**: 面粉的量（克） * **x2**: 酵母的量（克） * **x3**: 盐的量（克） 而最终的输出（结果）是披萨的口感，由一位“口感专家”在1到10的评分标准中进行评价（10为最佳）。 **通过学习DOE，您将能够：** * **辨识关键因素：** 确定哪些因素（如面粉、酵母或盐的用量）对披萨的口感有显著影响。 * **量化影响程度：** 了解这些因素中，哪个因素对口感的影响最大。 * **探究交互作用：** 评估不同因素之间的交互作用是否会影响面团的质量。换句话说，当改变一个因素（例如酵母的用量）时，是否需要同时调整另一个因素（例如面粉的用量）才能维持理想的口感。 * **确定最佳组合：** 找到每个因素的最佳用量，以达到最理想的口感。 * **进行预测：** 能够预测在改变一个或多个因素的组合时，披萨的口感会如何变化。例如，根据600克面粉、40克酵母和4克盐的组合，预测面团的口感评分。 总而言之，DOE 是一种**规划实验并利用统计分析来评估实验结果**的强大方法。
What is DOE and what are it's benefits? Let us assume that you are an owner of a pizza restaurant and of course you want to make the tastiest pizza, not in the world, let us be realistic, in your town. The quality of the pizza has a lot of factors, the pizza dough (flour, yeast and salt), the filling (sauce and cheese), also the oven temperature, the time in the oven and some other factors too. We will focus in our example on the recipe of the pizza dough. So we have the factors: x1 the amount of flour in gram, x2 the amount of yeast in gram and x3 the amount of salt in gram. The output is the taste of the pizza, let us say we have a taste expert, who judges the quality of the dough on a scale of 1 to 10, with 10 being best. *fig 2*.By using the design of experiments method we can know:Which of the factors (flour, yeast, or salt) affect the taste.Which of those factors has the highest effectIf the interaction between the examined factors affects the quality of the dough. In other words, if we change for example the amount of the yeast, should we also have to change at the same time the amount of the flour to keep the taste at its desired level.What is the best amount of grams for every factor to reach the best tasteAnd finally, we can predict what the taste would be, in case we change one factor or more. For example, if we mixed 600 g flour with 40 g yeast with 4 g salt, we can predict from this combination of values, what would be the taste of our doughDoe is a method to plan an Experiment and then evaluate this experiment with statistical analysis.
